A model system for mitochondrial biogenesis reveals evolutionary rewiring of protein import and membrane assembly pathways

Abstract

The controlled biogenesis of mitochondria is a key cellular system coordinated with the cell division cycle, and major efforts in systems biology currently are directed toward understanding of the control points at which this coordination is achieved. Here we present insights into the function, evolution, and regulation of mitochondrial biogenesis through the study of the protein import machinery in the human fungal pathogen, Candida albicans. Features that distinguish C. albicans from baker’s yeast (Saccharomyces cerevisiae) include the stringency of metabolic control at the level of oxygen consumption, the potential for ATP exchange through the porin in the outer membrane, and components and domains in the sorting and assembling machinery complex, a molecular machine that drives the assembly of proteins in the outer mitochondrial membrane. Analysis of targeting sequences and assays of mitochondrial protein import show that components of the electron transport chain are imported by distinct pathways in C. albicans and S. cerevisiae, representing an evolutionary rewiring of mitochondrial import pathways. We suggest that studies using this pathogen as a model system for mitochondrial biogenesis will greatly enhance our knowledge of how mitochondria are made and controlled through the course of the cell-division cycle.

Mitochondrial biogenesis is an essential aspect of the cell-division cycle and requires replication of the mitochondrial genome, synthesis of lipids to build new membranes, and the import of around 1,000 different proteins encoded in the nuclear genome (1, 2). The pathways for protein import have been defined and are mediated by a series of molecular machines in the outer and inner mitochondrial membranes, along with soluble factors in the intermembrane space and matrix (3, 4). One of these molecular machines, the translocase of the outer membrane (TOM) complex, forms a channel in the outer mitochondrial membrane that serves as the gateway for protein import into mitochondria. Recent work in the model yeast Saccharomyces cerevisiae has shown that phosphorylation of specific subunits of the TOM complex regulates the activity of the protein import channel. An elegant coordination of mitochondrial biogenesis with cellular metabolic needs was seen when glucose levels were high, with phosphorylation of TOM complex subunits coinciding with repression of mitochondrial activity (5, 6). S. cerevisiae responds to the presence of glucose through metabolic cycling, oscillating cycles of aerobic respiration and glycolytic activity. This process requires tight regulatory control of mitochondrial biogenesis cordoned into specific states of metabolic activity that are coordinated with the cell-division cycle (7–9).

Although S. cerevisiae has provided an excellent, genetically tractable model organism to determine how cellular metabolism is integrated with signaling pathways and networks, it is somewhat peculiar with regards to regulation of mitochondrial biogenesis. This yeast is subject to the Crabtree effect: It ferments glucose anaerobically, whether oxygen is available or not (10). Many of the genes required for anaerobiosis arose through a whole-genome-duplication (WGD) event, with duplicated genes modified to provide new functions or new signaling switches (10–13). During optimal, rapid growth in glucose, mitochondrial function is repressed, and transcriptional networks have been reorganized in S. cerevisiae to enable regulation of mitochondrial biogenesis by carbon source (14, 15).

In anticipation of work that soon will enable genetic and biochemical investigations of the integration of metabolic control globally with every cellular pathway, we sought to develop a yeast model in which mitochondrial function is not repressed during optimal growth and which does not show metabolic cycling. This model would provide a means to distinguish those control points that are linked specifically to the switching events in metabolic control. Candida albicans is a budding yeast that diverged from the lineage that gave rise to S. cerevisiae around 300 million years ago (16). The genome of C. albicans has been sequenced (17), and various tools have been developed for gene deletion, epitope tagging, and genetic studies in this organism (18–21). Unlike S. cerevisiae, C. albicans grows aerobically, and mitochondrial respiration is active during optimal growth in glucose (15, 22, 23). Thus, although mitochondrial activity is regulated through the course of the cell cycle in C. albicans, it is not subject to changes depending on the available carbon source (15, 23, 24).

Here we demonstrate that C. albicans does not undergo metabolic cycling and that mitochondria can be isolated from C. albicans and assayed for the import of mitochondrial proteins in vitro. Using this model system, we find several differences in the import reactions relative to S. cerevisiae, uncovering functional aspects relating to the mitochondrial sorting and assembly machinery (SAM) complex, an essential molecular machine that assembles β-barrel proteins into the mitochondrial outer membrane. We show that in C. albicans the core SAM complex comprises at least three integral membrane proteins: Sam35, Sam50, and a protein which we refer to as “Sam51.” We demonstrate that Sam50 and Sam51 play similar roles in the assembly of the TOM complex, a process that is much more rapid in C. albicans than in S. cerevisiae. Moreover, consideration of the protein import pathways for components of the electron transport chain in C. albicans provides examples of evolutionary rewiring of mitochondrial import pathways in related species, showing that the particular pathway taken into mitochondria was not hard-wired early in evolution; instead, rewiring is open to organisms to take advantage of new metabolic or gene-regulatory opportunities.

Results

Protein Import Machinery in Mitochondria from C. albicans.

The components of the protein import pathways as discovered in S. cerevisiae are highly conserved in C. albicans, and BLAST searches are sufficient to identify candidate homologs in most cases (Table S1). Proteins of overall similar size and domain structures were detected, as detailed in Table S1. Several distinguishing features of C. albicans also were observed, e.g., in respect to the TOM complex: C. albicans has no Tom70 and has extensions to the cytosolic domains of Tom6 and Tom7 (Table S1). With recent studies in S. cerevisiae showing that phosphorylation in the cytosolic domains of TOM subunits is crucial for regulation of TOM complex activity in response to changes in metabolic activity (5, 6), future analysis of sequence variations in S. cerevisiae and C. albicans could help show how these phosphorylation sites affect cell cycle and metabolism. In some cases hidden Markov models (HMM) proved essential for identifying homologs; for example, an ORF encoding a Tom5-related sequence previously had been unidentified and unannotated because of its small size and sequence divergence. The Mia40 homolog in C. albicans previously was known only as a Hap43-repressed gene, because it lacks the characteristic N-terminal transmembrane signature of the Mia40 from S. cerevisiae. Using an HMM search, we also found a Sam50-related protein in C. albicans with no homolog in S. cerevisiae (Table S1). In S. cerevisiae, Sam50, Sam35, Sam37, and Mdm10 form the SAM complex, a molecular machine required for the assembly of β-barrel proteins into the mitochondrial outer membrane (3, 4).

Oxidative Phosphorylation and Protein Import into C. albicans Mitochondria.

During growth in glucose, S. cerevisiae represses mitochondrial function through changes in gene expression and metabolite levels via the yeast metabolic cycle (8). Chemostat cultures of S. cerevisiae revealed a cyclic change in oxygen consumption (Fig. 1A), consistent with that seen previously (8, 9). Metabolic cycling was not observed with cultures of C. albicans: These grew at comparably high cell density but had a constant requirement for oxygen consumption (Fig. 1A). We isolated mitochondria from cultures of C. albicans to test their capacity for protein import (Fig. 1B). The model precursor protein Su9-DHFR is destined to the mitochondrial matrix and is translocated across the outer and inner membranes to be processed in two steps, thereby generating an intermediate (i) form and then a mature (m) form of the imported protein (25). The intermediate and mature forms were protected from trypsin by the intact mitochondrial membranes (Fig. 1B). Membrane potential (ΔΨ_m)-dependent import was seen for Su9-DHFR and the other matrix proteins ScF₁β, ScAdh3, and ScCoxVa (“Sc” denotes proteins from S. cerevisiae) (Fig. 1B). The precursor (p) form of these proteins binds to the surface of mitochondria from C. albicans but is not imported. Treatment of the isolated mitochondria with trypsin degrades the surface-located precursor forms of all four proteins (Fig. 1B). The import rates of some precursor proteins are so high that the reaction was complete within 2 min when assayed at 25 °C but could be slowed at lower temperature (4 °C) to demonstrate a time-dependent rate of import (Fig. 1C).

Fig. 1.

Protein import into mitochondria from C. albicans. (A) Chemostat culture of S. cerevisiae CEN.PK and C. albicans DAY185 strains. (B) Mitochondria isolated from wild-type C. albicans (50 μg protein per lane) were incubated with the indicated proteins at 25 °C for the indicated times (min), treated with trypsin, and then analyzed by SDS/PAGE and phosphorimage analysis. The precursor (p), intermediate (i), and mature (m) forms are indicated where relevant. (C) Mitochondria isolated from C. albicans (50 μg protein per lane) were incubated with the indicated proteins at either 25 °C or 4 °C for the indicated times (min), treated with trypsin, and then analyzed by SDSPAGE and phosphorimage analysis.

Stop-Transfer Pathway: Rewiring the Import of Cytochromes.

A distinguishing feature of mitochondria isolated from C. albicans is the inefficiency with which they import Cyb2, a heme-containing dehydrogenase (l-lactate cytochrome-c oxidoreductase) essential for the utilization of l-lactate as a carbon source. Fig. 2A details the pathway for Cyb2 import via the stop-transfer sorting pathway as it occurs in S. cerevisiae. The import of ScCyb2 depends on active unfolding of the heme-binding domain by the action of matrix-located Hsp70 (26–30). All the necessary components for the stop-transfer pathway are present in C. albicans (Table S1), but, as compared with S. cerevisiae, import of ScCyb2 is highly inefficient in mitochondria from C. albicans, with a greatly diminished processing to produce the intermediate (“iCyb2”) and mature (“mCyb2”) forms (Fig. 2B). Similarly, the import of cytochrome c₁ from S. cerevisiae (ScCyt1) also follows the stop-transfer pathway (31), and the conversion of pScCyt1 to iScCyt1 is less efficient in mitochondria from C. albicans (Fig. 2B).

Fig. 2.

The stop-transfer pathway for protein import into mitochondria. (A) The stop-transfer pathway occurs by the sequential interaction with the TOM complex (white rectangles) and the TIM23 complex (white triangles), and the sequential processing by the mitochondrial processing peptidase (MPP) in the mitochondrial matrix and the inner membrane peptidase (IMP) in the inner membrane. (B) Mitochondria isolated from C. albicans (50 μg protein per lane) were incubated with the precursor (p) forms of ScAdh3 or ScCyb2 or ScCyt1 at 25 °C for the indicated times (min), treated with trypsin, and then analyzed by SDS/PAGE and phosphorimage analysis. The intermediate (i) and mature (m) forms are indicated where relevant. (C) The CYB2 genes from S. cerevisiae and C. albicans were amplified and cloned into a plasmid for expression in a Δcyb2 strain of S. cerevisiae, and the transformed strains were monitored for growth in semisynthetic medium with either l-lactate or glucose as a carbon source. Wild type, blue; Δcyb2, red; Δcyb2 /ScCYB2, green; Δcyb2 /CaCYB2; purple. All strains grew to comparable density in glucose medium. (D) Sequence alignment of the N-terminal 28 residues of Cyb2 from C. albicans (CaCyb2), Cyb2 from S. cerevisiae (ScCyb2), and related yeasts. Sequence identity and similarity are indicated by asterisks and dots, respectively. The processing site that generates the mature form of ScCyb2 is the peptidyl bond between residues L84 and D85, four residues upstream from Q89.

The ORFs encoding CaCyb2 and ScCyb2 were cloned for expression from a plasmid, and Δcyb2 mutants of S. cerevisiae were transformed. On minimal growth medium with l-lactate as a carbon source, the Δcyb2 mutants depend on ScCyb2 activity to feed electrons from l-lactate into the electron transport chain (Fig. 2C). CaCyb2 could not support growth of the Δcyb2 mutant on l-lactate, although the transcript was expressed at high levels (Fig. S1A), and the functional domains of Cyb2 are highly homologous (Fig. S1B), suggesting that the Candida protein was not imported sufficiently well into the mitochondrial intermembrane space in S. cerevisiae. The import signals for delivery of the cytochromes Cyb2 and Cyt1 to the intermembrane space in S. cerevisiae are well characterized N-terminal extensions, and comparative sequence analysis shows that CaCyb2 and the Cyb2 homologs from related yeasts entirely lack this N-terminal stop-transfer sorting sequence (Fig. 2D). Similarly, CaCyt1 lacks a large segment from within the N-terminal stop-transfer sorting sequence found on ScCyt1 (Fig. S1B).

Sam35, Sam37, and the Assembly of Voltage-Dependent Anion-Selective Channel Oligomers in C. albicans.

The assembly of β-barrel proteins into mitochondria can be measured readily in vitro using blue-native PAGE (BN-PAGE), and two model proteins have been used as substrates in these assays: Tom40 and the voltage-dependent anion-selective channel (VDAC), also known as “porin.” In the case of Tom40, the assembly reaction measures both the assembly of the Tom40 β-barrel and the recruitment of other TOM subunits to form the mature TOM complex. When mitochondria isolated from C. albicans are used, the assembly of Tom40 from S. cerevisiae occurs at a rate similar to that of the assembly of Tom40 from C. albicans, demonstrating that there is sufficient sequence similarity in the interactive surfaces so that the heterologous Tom40 can be assembled into the oligomeric TOM complex (Fig. 3A).

Fig. 3.

Sam37 is a major determinant of SAM function in C. albicans. (A) Mitochondria (50 μg protein per lane) from C. albicans were isolated and assayed for import of [³⁵S]-CaTom40 and [³⁵S]-ScTom40, and were monitored by BN-PAGE and phosphorimage analysis. (B) Mitochondria were isolated from S. cerevisiae and C. albicans, assayed for import of ScVDAC (Por2), and analyzed by BN-PAGE (Upper) or SDS/PAGE (Lower), followed by phosphorimage analysis. (C) The expression level of SAM35 in the indicated strains was monitored by quantitative PCR. Shown are averages plus SE of three independent biological repeats assayed in duplicate (39). (D) Mitochondria from the same strains of C. albicans were isolated and assayed for import of [³⁵S]-CaTom40 for 30 min, visualized by BN-PAGE and phosphorimaging. The relative size of the SAM complex can be inferred from the migration of “intermediate I” (IntI), indicated by the asterisks. (E) Mitochondria from C. albicans wild-type, sam37ΔΔ, or sam37ΔΔSAM35↑↑ strains (50 μg protein per lane) were incubated with VDAC at 25 °C for the indicated times (min). Then samples were analyzed by BN-PAGE (Upper) or SDS/PAGE (Lower), followed by phosphorimage analysis. (F) Mitochondria from the indicated strains of C. albicans (50 μg protein per lane) were incubated with ScF₁β or ScAdh3 at 4 °C or withScAac1 at 25 °C for the indicated time (min), treated with trypsin, and then analyzed by SDS/PAGE and phosphorimage analysis.

The major β-barrel protein in the outer mitochondrial membrane is the homo-oligomeric VDAC. Newly imported VDAC monomers assemble into oligomers with preexisting VDAC (32). ScVDAC is inserted into the outer membrane in a time-dependent manner forming numerous oligomers, including a major 200-kDa form, which are resolved by BN-PAGE (Fig. 3B). In contrast, the oligomers of ScVDAC in C. albicans mitochondria are of uniform stoichiometry and/or are less dynamic in nature: The imported ScVDAC showed the majority of the protein in stable oligomers of ∼400 and ∼440 kDa (Fig. 3B). In S. cerevisiae, the import and assembly of VDAC depends largely on the SAM complex subunit Sam37 (33). In both S. cerevisiae and C. albicans, deletion of Sam37 leads to a concomitant reduction in levels of Sam35, and this reduction can be suppressed by overexpression of the SAM35 gene (34, 35), With a sam37ΔΔ strain of C. albicans engineered to have one copy of the SAM35 gene under the control of the strong TEF1 promoter, a 9.5-fold up-regulation of SAM35 transcript levels was measured in the mutant (sam37ΔΔSAM35↑↑) (Fig. 3C). That Sam35 then is assembled into the SAM complex in the sam37ΔΔSAM35↑↑ strain is evident by the change in the size of Tom40 assembly intermediate I compared with the sam37ΔΔ strain (Fig. 3D). Only a minor defect in VDAC assembly is seen in mitochondria from the C. albicans sam37ΔΔ strain, and the increased levels of SAM35 did not rescue this defect (Fig. 3E). The very mild phenotype seen in C. albicans is in contrast to the strong requirement for Sam37 in VDAC assembly measured in S. cerevisiae (33). For quality control, we show that mitochondria from the sam37ΔΔSAM35↑↑ strain import ScF₁β, ScAdh3, and ScAac1 with efficiency comparable to that of mitochondria from the wild type strain under conditions where protein import was linear with respect to time and dependent on the membrane potential (ΔΨ_m) (Fig. 3F).

TOM Complex Assembly in C. albicans.

In S. cerevisiae, the β-barrel core of the TOM complex, Tom40, is assembled via two intermediate forms into the mature TOM complex (Fig. 4A) in a time-dependent manner that, in vitro, takes more than 60 min to complete (34, 36). Mitochondria from C. albicans assemble Tom40 into the mature TOM complex much more rapidly: When [³⁵S]-labeled CaTom40 was imported into mitochondria isolated from C. albicans, Intermediate I (representing [³⁵S]-CaTom40 bound in the SAM complex) forms within 2 min and is transferred into the mature, ∼400-kDa form of the TOM complex within 10 min (Fig. 4B). The loss of Sam37 has a major impact on the assembly of the TOM complex, including a decrease in the amount of [³⁵S]-CaTom40 entering the intermembrane space as judged by SDS/PAGE and BN-PAGE. These effects are suppressed partially by overexpression of SAM35 (Fig. 4B). These results in C. albicans are consistent with the functions of Sam35 and Sam37 observed in the S. cerevisiae model system (34, 37).

Fig. 4.

Sam35 is an integral membrane protein functioning in substrate entry into the SAM complex. (A) Schematic of the assembly pathway for the Tom40 precursor, showing its incorporation into assembly intermediate I, assembly intermediate II, and the mature TOM complex. (B) Mitochondria (50 μg protein per lane) from the indicated strains of C. albicans were isolated and assayed for the import of CaTom40. Equal samples were withdrawn for analysis by BN-PAGE (Upper) or SDS/PAGE (Lower) and phosphorimaging. (C) Hydrophobicity analysis of CaSam35 with dense alignment surface method (DAS) showed it to have a more hydrophobic character than the prototypical ScSam35 and a conserved DUF2731 domain (35) indicated by the pink shape, with the gray bar representing the protein drawn to scale with the x axis. Schematics of HA-tagged versions of Sam35 are shown also. Mitochondria from C. albicans HA-SAM35 strain and SAM35-HA were isolated, incubated with proteinase K for 20 min at 4 °C, and then analyzed by SDS/PAGE and immunoblotting. Tom71 is an outer membrane protein. Tim23 is an inner membrane protein. (D) Mitochondria from C. albicans HA-SAM35 were isolated and incubated with 0.1 M Na₂CO₃ (pH 11.5), and the extracted membranes were reisolated by centrifugation. The extract supernatant (“extract”) and the extracted membranes (“membrane”) were analyzed by SDS/PAGE and immunoblotting. Tim10 is a soluble protein in the intermembrane space. (E) Mitochondria from C. albicans SAM35-HA were isolated and incubated with 0.1 M Na₂CO₃ (pH 11.5), and the extracted membranes were reisolated by centrifugation and analyzed by SDS/PAGE and immunoblotting.

The Sam35 subunit of the SAM complex is enigmatic, because it interacts selectively with substrate β-barrel proteins in the intermembrane space but behaves as if it were a peripheral membrane protein on the outer surface of mitochondria: It is largely accessible to proteolysis and is readily extracted from mitochondrial membranes using sodium carbonate treatment (37). Hydrophobicity analysis of CaSam35 suggested that it has potential membrane-spanning segments that are more hydrophobic than the amphipathic segments detected in ScSam35 (Fig. 4C). We sought to take advantage of this difference and use the more hydrophobic C. albicans protein to determine whether Sam35 has a transmembrane segment. Mitochondria isolated from a strain expressing N-terminally epitope-tagged HA-CaSam35 and from a strain expressing C-terminally epitope-tagged CaSam35-HA were treated with trypsin, and in both cases CaSam35 behaves as an outer membrane protein, exposed on the mitochondrial surface and sensitive there to protease (Fig. 4C). Incubation of mitochondria with sodium carbonate to strip peripheral membrane proteins revealed that both HA-CaSam35 (Fig. 4D) and CaSam35-HA (Fig. 4E) behave as an integral membrane protein.

C. albicans has a Second Member of the Omp85 Protein Family.

The feature distinguishing SAM complex components in C. albicans from those in S. cerevisiae is a second protein sharing 22.8% sequence identity and predicted domain composition with ScSam50 (Fig. S2A). We refer to this protein as “Sam51.” This protein is not simply a result of gene duplication from the well-characterized WGD event, because we found homologs of Sam51 in most clades of yeast, including species prior and subsequent to the WGD event (Fig. S2B). Sam51 is found throughout the Ascomycotina, indicating that its absence in S. cerevisiae is the result of a secondary loss in this particular species (Fig. S2C and Fig. 5). Reconstruction of the ancestral lineage at the point of genome duplication predicts that the ancestor had both SAM51 and SAM50 genes, with data for the ancestor derived from the Yeast Gene Order Browser (38). In the post-WGD species, there are various permutations in the observed losses of the SAM51 and SAM50 genes (Fig. S2C). The Sam51 proteins are as similar to bacterial BamA proteins as to the Sam50 sequences, as shown by the clear branching pattern forming three independent subgroups of the Omp85 protein family (Fig. 5). This phylogenetic analysis demonstrates that the Sam51 group of sequences forms a monophyletic branch with the exclusion of all Sam50 sequences and therefore represents a separate subgroup of Omp85 proteins (Fig. 5).

Fig. 5.

The Omp85 protein family in C. albicans. (A) Phylogenetic analysis of Sam50, Sam51, and bacterial BamA sequences (Materials and Methods). The tree demonstrates that Sam51 sequences share a common ancestor to the exclusion of all Sam50 sequences as well as to the bacterial BamA sequences, recovered as the subgroup most similar to Sam51 based on BLASTP scores. Shown in blue are species that have both Sam50 and Sam51 genes; species with Sam50 but a secondary loss of Sam51 are indicated in red. Accession numbers for all sequences are given in Fig. S2D. (B) RT-PCR showing expression levels of SAM50 and SAM51 in C. albicans. PCR products from three independent cultures grown in rich medium (YPD) to log phase, with the linear range for each primer pair determined by assaying the appearance of the PCR products in time at different cycles. Shown are products that fell within the linear range of the reaction. (C) Purified Sam50 and Sam51 (0–100 ng protein) were used as standards alongside samples of mitochondria (25, 50, and 100 μg mitochondrial protein) analyzed by SDS/PAGE, immunoblot analysis, and phosphorimage quantitation to estimate the amount of Sam50 and Sam51 in 100 μg mitochondria.

RT-PCR shows that under standard growth conditions the SAM51 gene is expressed at higher levels than SAM50 in wild-type C. albicans (Fig. 5B). We raised antibodies to recombinant Sam50 and Sam51 and used a semiquantitative immunoblot, calibrated with titrations of purified Sam50 and Sam51, to determine that the steady-state protein levels of Sam50 and Sam51 are similar in C. albicans mitochondria (Fig. 5C).

SAM51 is not essential, and a homozygous diploid mutant sam51ΔΔ was created from which mitochondria were isolated. Tom40 is imported and assembled rapidly in mitochondria from C. albicans (Figs. 3A and 4B). To analyze the kinetics of assembly of the TOM complex in the absence of Sam51, [³⁵S]-CaTom40 was bound to mitochondria from wild-type C. albicans or the sam51ΔΔ mutant and was chased into the TOM complex (Fig. 6A) or was monitored in standard assembly assays run at 16 °C (Fig. 6B). Both assay systems showed that Sam51 plays a role in assembly of the TOM complex. Although the [³⁵S]-Tom40 substrate occupied the “assembly intermediate I” stage in mitochondria from wild-type cells, the assembly intermediate is somewhat more transient in mitochondria lacking Sam51 (Fig. 6A). The assembly of VDAC proceeded so rapidly, even at 16 °C, that no defect was observed in the kinetics of assembly (Fig. 6C), but a small decrease in the steady-state level of VDAC was seen in mitochondria from the sam51ΔΔ mutant (Fig. 6D). Thus, although Sam51 plays a role in β-barrel assembly, deletion of Sam51 results in only a mild defect of the assembly process.

Fig. 6.

Sam50 and Sam51 function in outer membrane protein assembly. (A) Mitochondria (50 μg protein per lane) from C. albicans wild-type or the sam51ΔΔ strains were incubated with CaTom40 for 10 min at 25 °C, isolated, and resuspended in import buffer at 25 °C. Samples were taken after the indicated time of chase for analysis by BN-PAGE and phosphorimaging. (B) Mitochondria (50 μg protein per lane) from C. albicans wild-type or sam51ΔΔ strains were isolated and assayed for import of CaTom40 at 16 °C and analyzed as above. (C) Mitochondria (50 μg protein per lane) from C. albicans wild-type or the sam51ΔΔ strains were isolated and assayed for import of VDAC at 25 °C or 16 °C. At the indicated times (min), a sample was removed for analysis by BN-PAGE and phosphorimaging. (D) Complete shutdown of SAM50 expression in the sam50∆SAM50↓↓ strain was verified by quantitative PCR. RNA was extracted from cells grown for either 3 or 24 h in the presence of 2.5 mM methionine and 0.5 mM cysteine to repress the MET3 promoter. Mitochondria were isolated from the C. albicans sam50ΔSAM50↓↓ strain after 3 h incubation in repressive conditions for comparison with wild-type and the sam51ΔΔ strain using SDS/PAGE and immunoblotting with the indicated antisera. (E) Growth of the C. albicans sam50ΔSAM50↓↓ strain on permissive (without methionine and without cysteine) and repressive (2.5 mM methionine/0.5 mM cysteine) medium. Growth at 30 °C was assessed after 3 d incubation. (F) The C. albicans sam50ΔSAM50↓↓ strain and the sam51ΔΔ strain were grown in parallel liquid cultures and shifted to repressive conditions for the final 3 h of incubation. Mitochondria then were isolated and assayed for import of CaTom40 at 25 °C. At the indicated times (min), a sample was removed for analysis by BN-PAGE and phosphorimaging. The position of the assembly intermediates I and II and the mature TOM complex are indicated.

Attempts to recover a homozygous diploid mutant sam50ΔΔ were unsuccessful, consistent with an essential role for the SAM50 gene in viability of C. albicans. A strain was engineered with SAM50 gene expression under the control of a repressible promoter, and growth on agar under repressive conditions showed a nearly total loss of viability (Fig. 6E). After only 3 h in liquid culture, gene expression was shut down in this sam50ΔSAM50↓↓ strain, with transcriptional repression confirmed by quantitative PCR (39), so that expression was below the level of detection (Fig. 6D). Mitochondria were isolated after growth under repressive conditions for 3 h, and Western blots using antisera specific for either Sam50 or Sam51 show there is no compensatory up-regulation of Sam51 levels in the sam50ΔSAM50↓↓ strain (Fig. 6D). Mitochondria isolated from the sam50ΔSAM50↓↓ strain and the sam51ΔΔ strain grown on the equivalent minimal medium were assayed for TOM complex assembly and showed that the sam51ΔΔ mutant and the sam50ΔSAM50↓↓ mutants have equivalent rates and extent of assembly of the [³⁵S]-Tom40 subunit into the mature TOM complex (Fig. 6F). We conclude that both Sam50 and Sam51 participate in the assembly of Tom40 and have overlapping roles in the entry and exit of substrate proteins through the SAM complex.

Discussion

Characterization of mitochondrial protein import in C. albicans opened the way to using this yeast as a model for mitochondrial biogenesis. Distinctions in the protein import pathway in C. albicans include an evolutionary rewiring of mitochondrial protein import pathways between related species. Furthermore, the C. albicans system revealed aspects of the function of the mitochondrial SAM complex that can be exploited in experiments directed at understanding the role that each component plays in the assembly of the TOM complex and other proteins in the mitochondrial outer membrane.

Evolutionary Rewiring of Targeting Routes.

It is well documented that S. cerevisiae has coevolved a forceful import motor and bipartite targeting sequences to deliver cytochromes into the intermembrane space via the stop-transfer pathway (reviewed in refs. 3 and 4). C. albicans has the capacity for import by the stop-transfer pathway, as evidenced by cytochrome c₁ from S. cerevisiae reaching the intermembrane space in C. albicans mitochondria and being processed there (Fig. 2B). Surprisingly, the cytochromes c₁ and b₂ in C. albicans and related yeasts lack the necessary targeting sequences to take the stop-transfer route into the mitochondrial intermembrane space, suggesting an evolutionary rewiring of targeting routes. Consistent with this model, the C. albicans CYB2 gene could not rescue the S. cerevisiae ∆cyb2 mutation, although it was expressed at high levels and the functional domains of Cyb2 are highly conserved between these two yeasts. How then do the cytochromes enter the mitochondrial intermembrane space in C. albicans? Three potential routes are available: (i) the disulfide-relay pathway, mediated by Mia40; (ii) the cytochrome c pathway, mediated by heme lyases; or (iii) another route, not previously described.

Both CaCyb2 and CaCyt1 have several cysteine residues that in principle could allow for interaction with Mia40. Although these residues are not the CX₃C or CX₉C motifs characteristic of most MIA substrates (3, 4), emerging data indicate that substrates that have CX₂C and other arrangements of cysteine residues (e.g., Erv1, Ccs1) use the MIA pathway for import into the intermembrane space (40–42). An alternative pathway would be driven by cytochrome heme lyases (e.g., Cyt2 and Cyc3), by analogy with the pathway for cytochrome c that is driven by the heme lyase Cyc3. In this import pathway the attachment of heme can serve to drive completion of polypeptide transfer across the outer membrane (43, 44). The proposition that evolution has rewired targeting routes might provide a broad explanation to some curious past examples in mitochondrial protein import in diverse organisms. For example, the cytochrome c₁ found in the protozoan Trypanosoma brucei has no recognizable stop-transfer sequence, and in protist mitochondria TbCyt1 is imported into mitochondria in the absence of a membrane potential, leading to the suggestion that a route other than stop-transfer is used for its entry into the intermembrane space (45). Our results also shed light on differences observed between S. cerevisiae and humans. In S. cerevisiae the Mia40 protein is imported via the stop-transfer pathway, but in humans the homolog of Mia40 is imported via the Mia40 pathway (3, 4). This difference had been puzzling, initially calling into question the functional homology between the yeast and human Mia40 proteins. Based on our data for the C. albicans cytochromes, for which the functional and structural conservation with S. cerevisiae cannot be doubted, we suggest that the distinct import pathways are not an inevitable consequence of an event in the earliest eukaryotes that has been hard-wired into mitochondrial biogenesis but instead reflect rewiring options open to organisms to take advantage of new evolutionary opportunities.

Assembly of VDAC and Assembly of the TOM Complex.

VDAC is required for metabolite exchange between the cytosolic and mitochondrial pools. In S. cerevisiae VDAC exists in numerous packing densities and oligomeric forms highly sensitive to metabolic conditions in the cell, ranging from monomers and trimers to regions of membrane where the packing density is 80% VDAC, functioning as a voltage-dependent molecular sieve (46). In S. cerevisiae the assembly of these dynamic VDAC oligomers is highly sensitive to the presence of Sam37 (33). The situation in C. albicans is quite distinct, with an apparently more regular arrangement of VDAC in a predominant oligomeric form and with little effect observed for Sam37 in the assembly of this VDAC oligomer.

Sam35 is necessary for newly imported [³⁵S]-Tom40 to engage productively with the SAM complex (34, 37) and functions as the receptor for the targeting signal present at the C terminus of Tom40, VDAC, and other β-barrel proteins (37). Until now, it has been difficult to rationalize how Sam35, thought to be a peripheral protein on the outer surface of the outer membrane, could perform this function on the inner surface of the outer membrane. The topology of CaSam35 explains this apparent contradiction: Although most of Sam35 is exposed to the cytosol, at least one transmembrane span is integrated into the outer membrane, leaving some part of Sam35 exposed in the intermembrane space and thus explaining how it can function as a receptor for β-barrel proteins. Sodium carbonate extraction is an empirical method that depends on the hydrophobicity of a transmembrane span to resist chaotropic forces at pH ∼11 (47), and the relatively amphipathic nature of ScSam35 makes this integral membrane protein in this species prone to alkali extraction from the mitochondrial outer membrane.

Although at present Sam 51 has been described only in C. albicans, molecular phylogenetics (Fig. S2B) and comparative genome analysis (Fig. S2C) demonstrate that the presence of genes encoding both Sam50 and Sam51 is the ancestral yeast condition. In the course of evolutionary time, through the WGD event, sam51 genes were lost from some yeasts, such as S. cerevisiae. Given that Sam51 seems to be dispensable, it is of great interest that none of the yeasts in the same category as C. albicans have, in fact, dispensed with its function. In C. albicans, assembly of newly imported [³⁵S]-Tom40 into the TOM complex is much faster than in S. cerevisiae and is driven by both Sam50 and Sam51. Mitochondria with a nearly complete depletion of Sam50 or complete depletion of Sam51 have similarly mild defects in TOM complex assembly. This result is consistent with Sam50 and Sam51 playing equivalent roles in β-barrel assembly but is in sharp contrast to the growth phenotypes of the two mutant strains, with a loss of viability seen only for the deletion of the SAM50 gene. Given the emergence of data from S. cerevisiae suggesting the involvement of subunits of the SAM complex in contacts between the mitochondrial outer membrane and endoplasmic reticulum (48, 49) and between the mitochondrial outer membrane and the mitochondrial inner membrane (50), C. albicans is a unique system in which to address the essential function of Sam50 that might distinguish it from Sam51.

Concluding Remarks.

Taken together, the rewiring of protein import pathways, the less dynamic use of VDAC for metabolite exchange, and the subtle differences in the subunits contributing to the activity of the SAM complex show the utility of studying the mitochondrial import machinery in C. albicans. We anticipate that parallel investigations into the regulation points of TOM complex assembly and mitochondrial protein import using both C. albicans and S. cerevisiae will provide a more comprehensive picture of import pathway mechanisms and of the links among metabolism, the cell cycle, and mitochondrial biogenesis.

Materials and Methods

Yeast Strains.

The C. albicans strains are derivatives of BWP17 (51). The wild-type strains were DAY185 (URA3⁺ ARG4⁺ HIS1⁺) and DAY286 (URA3⁺ ARG4⁺) (52). The C. albicans sam37ΔΔ mutant and the sam37ΔΔ mutant overexpressing SAM35 (sam37ΔΔ SAM35↑↑) under the strong constitutive promoter TEF1 have been described previously (35). The sam51ΔΔ mutant was constructed by standard methods using PCR and homologous recombination, with the URA3 and ARG4-based selection cassettes (48). The conditional C. albicans sam50 mutant (sam50∆SAM50↓↓) was made by deleting the first allele with the ARG4 marker cassette and placing the other allele under the control of the repressible MET3 promoter (1,362 bp of the 5′ UTR of MET3) by constructing a URA3-MET3 promoter fusion cassette. The MET3 promoter was repressed by the addition of 2.5 mM methionine and 0.5 mM cysteine to synthetic growth medium. The epitope-tagged Sam35 strain was constructed by fusing a single HA tag to the N terminus of one of the alleles of SAM35 under the TEF1 promoter using the plasmid pCJN498 (53). The C-terminal fusion of the HA tag was performed as described in ref. 21. For complementation of the S. cerevisiae Δcyb2 mutation, the CYB2 gene was deleted in the YPH499 strain background (MATa ade2-101 his3-200 leu2-1 ura3-52 trp1-63 lys2-801), and the mutant was transformed with either the empty parental plasmid (2µ/URA3) or a plasmid containing either ScCYB2 or CaCYB2 driven by the ADH1 promoter. Growth was monitored in synthetic medium lacking uracil at 30 °C, with either 2% (wt/vol) glucose or 2% (wt/vol) l-lactate (pH 4.6) as the carbon source.

For metabolic cycling experiments prototrophic strains of yeast were grown as previously described (8). The S. cerevisiae strain CEN.PK and the C. albicans strain DAY185 (52) were used in these experiments.

Sequence Analysis.

The methodology for HMM analysis has been described previously (54), and HMMs were used to scan UniProt (Release 12.4, containing Swiss-Prot Release 54.4 and TrEMBL Release 37.4). Searches of the C. albicans genome refer to the type strain SC5314 and Assembly 21 of the genome information. The C. albicans Sam51 sequence was identified originally by HMM searches, and further homologs were identified by BLASTP and are available in Fig. S2D. Alignments were calculated using MUSCLE (55) under the default settings, and poorly conserved sites were removed manually. The best-fitting rate exchange matrix was determined with ProtTest under the Akaike information criterion (AIC) and resulted in LG + G for both datasets; tree calculations were performed with PhyML 3.0 (56) with the LG rate exchange matrix (57), tree topology search was performed with the Best of NNIs and SPRs option, 500 bootstrap calculations, and all other settings as default. Bootstrap support is shown as percentage values.

Preparation of Mitochondria.

For the preparation of mitochondria, cultures of S. cerevisiae strain W303 were grown in rich medium [YPAGal; 2% (wt/vol) yeast extract, 1% peptone, 0.01% adenine, 2% (wt/vol) galactose]. C. albicans strains were grown in rich yeast extract peptone dextrose (YPD) medium [2% (wt/vol) glucose] at 30 °C. Mitochondria were isolated by differential centrifugation based on the method of Daum et al. (58). Homogenization was performed in 0.6 M sorbitol, 20 mM K⁺ Mes (pH 6.0), 1 mM PMSF. The mitochondrial pellet was resuspended in 0.6 M sorbitol, 20 mM K⁺ Hepes (pH 7.4), 10 mg/mL BSA.

Protein Import Assays and Electrophoresis.

In vitro translation and import assays were performed as described (34). When necessary, the membrane potential was dissipated by the addition of a 100× solution of antimycin (8 μM), valinomycin (1 μM), and oligomycin (20 μM) to mitochondrial samples. Proteins were analyzed by SDS/PAGE or by BN-PAGE as previously described (34).

Antibody Production.

Open-reading frames encoding Sam50 and Sam51 were synthesized in codon-optimised form for expression in Escherichia coli (GenScript) and cloned into a pET9 expression vector with an N-terminal 10×His tag. Recombinant expression in E. coli produced inclusion bodies containing the proteins. This material was purified by washing with 10% (wt/vol) TritonX-100 in PBS and using Ni-NTA agarose (Qiagen) as per manufacturers instructions, and the material was emulsified for immunization of rabbits. Other antisera used were raised against mitochondrial proteins from S. cerevisiae and cross-react with the homologous protein in C. albicans.